This invention relates to integrated and guided-wave optical devices, including intensity and phase modulators.
More particularly, the invention relates to annealed proton-exchanged integrated optic chip (IOC) devices which each utilize a LiNbO3 of LiTaO3 substrate.
In a further respect, the invention relates to a process which greatly improves the DC drift problem commonly associated with LiNbO3 and LiTaO3 IOC devices, and which also reduces the insertion loss and improves the fabrication yield of such devices.
A LiNbO3 crystal or substrate has several properties that are useful in photonics. The most useful property is the electro-optic (EO) effect. When an electric field is applied to the LiNbO3 crystal, the refractive index is changed according to the strength of the electric field. If and optical wave is traveling through the crystal, the electric field also modulates the phase of the optical signal. The ability of the electric field to modulate the phase of the optical signal is called the electro-optic (EO) effect. The EO effect is, however, limited. To take advantage of the EO effect, a compact waveguide structure is formed on the LiNbO3 (lithium niobate) crystal.
A lithium niobate crystal is processed in a wafer form in a manner similar to that utilized to process silicon crystals or wafers. A waveguide is formed in the lithium niobate crystal simply by increasing the refractive index in selected areas of the crystal. The waveguide ordinarily is only a few microns wide and thus is compatible with the dimension of an optical fiber.
Integrated optic chips (IOCs) were developed in the 1980s. Thousands of IOCs have been installed in CATV transmission equipment, digital telecommunications network equipment, fiber-optic gyros (FOGs), optical switches, and optical sensors. Advantages of IOCs include low-optical loss, low-voltage drive, high-frequency bandwidth, and small size and weight. Various types of IOCs are illustrated in FIG. 1.
Two prior art methods are utilized to fabricate LiNbO3 devices. Both methods utilize conventional photolithography and vacuum deposit techniques found in semiconductor processing.
The first prior art method for fabricating LiNbO3 devices is the titanium in-diffusion process. During this process a titanium metal strip is deposited on a LiNbO3substrate. The substrate and titanium strip are heated to 1050 degrees C. to diffuse titanium into selected portions or areas of the substrate to form waveguide channels in the substrate. Titanium ions diffuse to interstitial sites in the substrate and function collectively as an impurity dopant. The diffusion of titanium ions into the substrate increases the refractive indices of the substrate. The process was developed in the early 1070s.
The second prior art method is the proton exchange process. During this process a masked LiNbO3 substrate is immersed in a molten bath of pure benzoic acid at a temperature below the decomposition point of the acid. The acid causes lithium ion at or near the surface of unmasked areas of the LiNbO3 substrate to be replaced by hydrogen ions from the acid bath. Replacement of Li+ ions with of H+ ions increases the refractive index of the unmasked areas of the substrate to form a waveguide channel(s) in the substrate. The proton exchange process was developed in the early 1980s and improved in the late 1980s.
The proton exchange process has major drawbacks. First, the process ordinarily is carried out with molten benzoic acid. Benzoic acid is a solid crystallized substance at room temperature. Consequently, the benzoic acid must be heated to a temperature over 200 degrees C. in order to carry out the proton exchange process. At this temperature, the vapor pressure of the molten benzoic acid is high and the process can become unstable.
In addition to the processing problems noted above, treatment of a LiNbO3 substrate with benzoic acid produces severe crystal structure damage in the substrate. The substrate must be annealed to reduce the damage. Problems caused by crystal damage include high propagation loss due to scattering, instabilities in the refractive index of the waveguide channels, and degradation of the electro-optic effect. Annealing the substrate minimizes these problems. One of the most important of such problems is DC stability. When an annealed IOC substrate (produced using a benzoic acid bath) is utilized for the Mach-Zehnder Interferometer (MZI) or switched, the IOC substrate may need to be biased with an electric voltage to operate properly. For example, in the CATV modulator application, a MZI needs to be biased at the so-called xe2x80x9cquadraturexe2x80x9d (xcfx80/2) point to achieve maximum linearity. For a switch application, a bias voltage is needed to maintain the device in minimum of maximum light transmission.
After a LiNbO3 substrate is processed in a molten benzoic acid bath, the residual crystal damage and mobile charge migration present in the substrate cause the present bias point for the IOC to drift from the original set point of the xcfx80 phase over a period of time (See FIG. 2). This drift occurs at a very low frequency (0.001 Hz to 0.0001 Hz) and is called DC drift. The period of time which passes before such drift occurs can be as short as a few minutes or as long as several hours. Once drift occurs, the MZI is biased at an incorrect point and the modulated signal includes severe non-linearity. The DC drift may eventually even move beyond the maximum supplied voltage from the power supply. If this occurs, the entire transmission system must be reset in order to operate. Such DC voltage drift can therefore seriously compromise the integrity of the MZI.
DC drift is believed to occur because of the mobile charge and ions that are produced during treatment of a LiNbO3 substrate in a benzoic acid bath. When a bias voltage LiNbO3 substrate is applied to the substrate, the mobile charge and ions in the waveguide region slowly move toward opposing polarity electrodes and set up a counter electric field which cancels the applied E-field from the electrode. This is illustrated in FIG. 3.
The severity or magnitude of DC bias drift can vary among substrates processed in a molten benzoic acid bath.
One solution to the DC drift problem is providing a mechanism for resetting the bias point with a refreshing voltage directed to the bias electrodes of the IOC. The refreshing voltage restores the IOC to its original setting. The utilization of a refreshing voltage, however, requires the use of a monitoring system to measure the drift and initiate the necessary resetting sequence. This makes the system design much more complicated and renders the IOC useless for many applications.
In the past, sulfuric acid has sometimes been utilized to treat LiNbO3 substrates. The use of sulfuric acid has not been preferred because it resulted in high loss and lower yield of serviceable substrates.
Accordingly, it would be highly desirable to provide an improved IOC substrate and method for making the same which would eliminate or minimize DC drift in the IOC.
Therefore, it is a principal objective of the invention to provide an improved IOC and method for making same.
A further objective of the invention is to provide an improved IOC in which DC drift is minimized or is substantially eliminated so that a modulator or switch can be produced which can maintain its bias for extended periods of time.